The purpose of this research proposal is to conduct a methodologically rigorous and thorough evaluation of a computer- automated ECG system currently being installed at the Northwestern University Medical School. Atlhough there is a growing research literature in this area, there are a number of cogent reasons for conducting such an evaluation. First, the characteristics of the site including those of the cardiology staff make it highly unlikely that a universal program will be acceptable. This is especialy the case at Northwestern which is a high-volume (over 40,000 ECGs/year) laboratoy where over 90 percent of the ECGs are abnormal. For this reason the USPHS ECAN D program has been extensively revised for use at this site. Second, methodologicaly sophisticated estimates of computer-physician and intra-physician reliability using double-blind procedures to eliminate various sources of bias have not been employed. And, third, the system will also be involved in the teaching curriculum of the medical school. The effects of such computer interpretations on the training of physicans has not been previously investigated. The proposal suggests three experiments employing the latest techniques in true and quasi- experimental design to deal with these issues in evaluating the Northwestern ECG computer system.